Linearization and state estimation of unknown discrete-time nonlinear dynamic systems using recurrent neurofuzzy networks

Model-based methods for the state estimation and control of linear systems have been well developed and widely applied, In practice, the underlying sy stems are often unknown and nonlinear, Therefore, data based model identifi cation and associated linearization techniques are very important. Local li nearization and feedback linearization have drawn considerable attention in recent years. Ln this paper, linearization techniques using neural network s are reviewed, together with theoretical difficulties associated with the application of feedback linearization, A recurrent neurofuzzy network with an analysis of variance (ANOVA) decomposition structure and its learning al gorithm are proposed for linearizing unknown discrete-time nonlinear dynami c systems. It can he viewed as a method for approximate feedback linearizat ion, as such it enlarges the class of nonlinear systems that can be feedbac k linearized using neural networks. Applications of this new method to stat e estimation are investigated with realistic simulation examples, which sho ws that the new method has useful practical properties such as model parame tric parsimony and learning convergence, and is effective in dealing with c omplex unknown nonlinear systems.